See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/325447054

Availability and reliability analysis of integrated reverse osmosis - Forward

osmosis desalination network

Article  in  Desalination and Water Treatment · March 2018

DOI: 10.5004/dwt.2018.22147

CITATIONS READS

14 1,563

5 authors, including:

Usman Safder Pouya Ifaei

UnU Inc. | Kyunghee University Kyung Hee University
32 PUBLICATIONS   361 CITATIONS    63 PUBLICATIONS   801 CITATIONS   

SEE PROFILE SEE PROFILE

Kijeon Nam Zhowun Rashidi

Kyung Hee University 30 PUBLICATIONS   462 CITATIONS   
43 PUBLICATIONS   477 CITATIONS   
SEE PROFILE
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Special Issue “The Nexus among Sustainable Development Goals and Clean Energies” View project

sustainable development View project

All content following this page was uploaded by Pouya Ifaei on 09 September 2019.

The user has requested enhancement of the downloaded file.

 Desalination and Water Treatment 109 (2018) 1–7
March
www.deswater.com
doi: 10.5004/dwt.2018.22147

Availability and reliability analysis of integrated reverse osmosis – forward

osmosis desalination network

Usman Safder, Pouya Ifaei, KiJeon Nam, Jouan Rashidi, ChangKyoo Yoo*
Department of Environmental Science and Engineering, College of Engineering, Center for Environmental Studies,
Kyung Hee University, Seocheon-dong 1, Giheung-gu, Yongin-Si, Gyeonggi-Do 446-701, Korea, Tel. +82-31-201-3824;
Fax: +82-31-202-8854; email: ckyoo@khu.ac.kr (C. Yoo)

Received 19 December 2017; Accepted 20 January 2018

abstract
The hybrid desalination of reverse osmosis (RO) with forward osmosis (FO) is an innovative technol-
ogy for freshwater production worldwide, which provides many advantages such as RO fouling and
scaling, recovery of the energy of RO brine, minimizing the power consumption and minimizing the
use of chemicals required for conventional pretreatment steps. For this reason, breakdowns, blockage
of the membrane, pressure losses and preventive maintenance have to be minimized in duration and
frequency to ensure the maximum availability and indeed improving the availability or reliability of
the system leads objectively to the reduction of operating costs. In this study, hybrid fault tree anal-
ysis method based fuzzy set theory was considered to evaluate the availability and reliability of the
integrated RO–FO desalination system. For assessing availability, the system was divided into four
subsystems such as seawater intake pump, pressurization stage, FO stage and RO stage. The data were
collected from reliability data bank and probabilities estimated by considering the fuzzy set theory,
wherever detailed data were not available. The overall unavailability of the hybrid system was 0.0156.
Keywords: Hybrid desalination system; Availability; Reliability; Fault tree analysis; Failure probabilities

1. Introduction distillation and vapor compression are the dominant

distillation technologies. Membrane processes such as reverse
The development of human life depends on the
osmosis (RO), electrodialysis and an emerging technology
availability of freshwater resources. Some regions have lim-
forward osmosis (FO) are used for different desalination
ited resources of freshwater while the others have abundant
purposes. The desalination technologies using membrane
resources. According to the study, more than one billion
require energy as mechanical or electrical. However, RO has
of people do not have access to freshwater [1]. The global
the largest installed capacity among them [4].
freshwater demand increases so that it is predicted that water
In the last few years, a growing interest in the application
scarcity is becoming new rising challenge [2].
of FO phenomenon in seawater desalination was observed.
Desalination is recognized as a necessary alternative
It is being investigated through laboratory and pilot scale
to overcome the water shortage dilemma [3]. Various
and has proved several advantages [5]. In FO, the flow of
desalination technologies were employed and are
water across the semi-permeable membrane occurs due to
classified into two categories: distillation and membrane
the osmotic pressure gradient between the two sides of the
technologies. Distillation converts saltwater into steam and
membrane. This process is established with no energy cost
then condenses into liquid. Distillation process consumes
for transmembrane flow except that required to circulate
thermal and electrical energy. Multi-stage flash, multi-effect
solutions in the system. The osmotic driving forces in FO can

* Corresponding author.

Presented at the 10th International Desalination Workshop (IDW 2017), November 22–25, 2017, Busan, Korea.
1944-3994/1944-3986 © 2018 Desalination Publications. All rights reserved.
2 U. Safder et al. / Desalination and Water Treatment 109 (2018) 1–7
be significantly greater than hydraulic driving forces in RO,
leading to higher water flux rates and recoveries [6].
The FO desalination process involves two steps. First,
the freshwater is extracted from the raw water source using
a draw solution while the second step deals with separa-
tion of the osmotic agent from the freshwater. Many options
were suggested including integration of FO with current
RO desalination process. The FO process was suggested
as a membrane pretreatment step for seawater prior to RO
desalination and has been used to desalinate feed streams
ranging from brackish water to saline [7]. Among several
hybrid FO systems, RO–FO hybrid desalination system,
which first proposed by Cath et al. [5] and later studied by
Kolluri et al. [8] to increase the product water quality and
to reduce the power consumption of seawater desalination.
Cost reductions are attributed to lower energy requirement
for pretreatment and RO process, fewer pretreatment pro-
cesses, lower maintenance and chemical costs and decrease
in RO membrane costs.
One of the important performance measures of any contin-
uously operating desalination system, apart from its product
water quality, power consumption and recovery ratio, is the
overall availability. Integrated RO–FO system was designed
for continuous operation; the availability of the system was
considered more appropriate than reliability and used as an
indicator of reliability. Improving the availability of the sys-
tem results in a frequent reduction in total production cost if
availability and reliability are maintained at high level.
During the last few decades, researchers around the world
have continued to make the improvement to the desalina-
tion processes as the quality of membranes, energy recovery
and combining various desalination processes to improve
the performance. On the other hand, only a few studies are
interested in improving the reliability and the availability
of desalination systems. Hajeeh and Chaudhuri [9] tried to
assess the availability of one RO plant operating in Kuwait
using fault tree analysis (FTA) method and showed that the
availability of the whole plant strongly depends on the plant
design and selection of subsystem’s material. Kutbi et al. [10]
applied the same approach (FTA) for reliability analysis of
RO plants operating in Jeddah, Saudi Arabia and showed
that the seawater intake system needs some improvements
and a proper preventive maintenance should be applied.
Bourouni [11] compared the two methods: FTA and reliabil-
ity block diagram (RBD) for availability assessment of one of
the RO plant operating in Kuwait and showed that the FTA is
too difficult to apply for complex systems.
Having reviewed the relevant literature, significant
effort has been devoted to performing the availability and
reliability study of RO plant. The reliability and availability of
the integrated RO–FO system should observe to improve the
system design and its operation. However, it is often difficult
to estimate precise failure probability of the components due to
vague characteristics of the events. Fuzzy methods might be the
only way when data are insufficient and vague. The fuzzy set
theory was introduced by Zadeh [12] to deal with the problem
in which the phenomena are imprecise and vague. The fuzzy
set theory provides a useful tool for directly working with the
linguistic expression in reliability analyses. Hence, the fuzzy
set concept was used to assess the failure rate of vague events
to estimate the probabilities of human unpredictability errors.
The main aim of the paper was to estimate the
availability and the reliability of the integrated RO–FO
network desalination system and contribute to assessing
the performance of the system, by utilizing FTA and fuzzy
sets. The goal was realized by achieving the following two
objectives:
• To identify the causes and sequences of failures in the
integrated RO–FO network.
• To assess the failure rate of the integrated RO–FO
network by identifying the components and events
causing downtime and unwanted effects.
2. Availability and reliability assessment of
desalination network
The objective of the availability and the reliability of any
continuously operating system is to maintain the desired
level of production on a continuous basis without failures
and to restore the system as each time it suffers from failure
[13]. Thus, management can achieve the minimum total cost
of production if availability and reliability are maintained at
a high level.
The reliability and availability of a system are systematic
approach used to identify and assess the frequency and
causes of failures. It reduces and controls the effects of fail-
ure to ensure satisfactory performance of the system [14].
Human errors and component failures do not only affect the
performance of the system but can also cause accidents. The
frequencies of such events are assessed during system design
and its operation. To enhance the performance and maximize
the profit of system, reliability analysis has to be done at the
design stage.
2.1. Availability and reliability
Reliability is defined as the ability of a device to perform
a required function under specified conditions during a
given period [15]. If the time-to-failure distribution indicates
the reliability (R(t)) of the integrated system and especially
exponential distribution with constant failure rate (ζ), the
reliability function of the steady state system is given by
Eq. (1):
R(t ) = e −ζt (1)
Maintainability is an important factor while designing
any continuously operating system. As the total downtime is
composed of the time for inspection and detection of faults,
then the time of repair faults and administrative time can be
minimized to reduce the total downtime. The maintainability
in terms of downtime is given in Eq. (2) as follows:
M(t ) = 1 − e − µt (2)
Here, M(t) represents the maintainability function and
µ represents the mean downtime in negative exponential
function. Since both reliability and maintainability are
important factors for continuously operating systems, these
two factors can be combined into one measure, availability.
 U. Safder et al. / Desalination and Water Treatment 109 (2018) 1–7 3

Availability is defined as the fraction of the time that

a component and system are able to perform its required
function [16]. The availability can be classified into two
distinct senses: interval availability such as the probability
that a system works on demand and inherent availability
such as the probability that the system is working at the
specific time. It can be used for continuously operating the
system and determined by Eq. (3) as follows:

O
A(t ) = (3)
O+D

Here, A(t) is the availability of the system, O and

D indicate the uptime and downtime, respectively.
Availability represents the probability that a system is in use
at a given time t interval and can be calculated using Eq. (4):

MTTF µ
A= =
MTTF + MTTR µ + ζ (4)

MTTR ζ
A= = (5)
MTTR + MTTF ζ + µ

where A in Eq. (5) represents the unavailability of the system.

MTTF is the mean time to failure and MTTR is the mean time
to repair and equivalent to (1/µ) and (1/ζ), respectively. In Fig. 1. Different methods for reliability and availability
above equations availability A is determined from failure assessment [17].
rate (ζ) and repairing rate (µ) assumed to be constant [17]
(exponential distribution). Both quantitative and qualitative information about the
system are obtained from FTA. The construction of FT provides
2.2. Availability modeling methods of industrial system a better understanding of the potential sources of failures,
which will lead to rethink the design and the operation per-
For analyzing the reliability of any designed and
spective to eliminate the potential hazards. FT can analyze the
operated system, many methods have been developed
system to identify the combinations of component failures,
such as FTA, failure mode effects and criticality analysis
operational errors, or other faults that initiate the top event.
(FMECA), RBD, reliability graph and Poisson’s process.
Finally, it may be used to calculate the demand failure proba-
Fig. 1 shows the different methods used in reliability and
bility, unreliability and unavailability of any system.
availability assessment [18]. Each method has its own pros
FT is a logical diagram that depicts the relationship
and cons. FMECA is qualitative analysis and reviewed many
between the events giving rise to the failure. “AND gates” are
components, assemblies and subsystems to identify the
used to connect the groups of events and conditions, if all of
failure modes, critical failures and their causes and effects.
them are required to be present simultaneously to cause the
Monte Carlo simulation draws a realization of each random
hazardous event to occur, which is shown in Fig. 2(a), while
variable and then estimates which units are down and for how
“OR gates” represents the existence of the alternative ways in
long. The system availability over the time can be determined
which a failure can occur, is shown in Fig. 2(b).
based on these calculations. Among all analytical techniques,
Consider the “AND” and “OR” FT as given in Fig. 2 where
FTA is the most important method used for reliability and
the simultaneous existence of the events E1, E2 … En results in
availability assessment of different kinds of systems.
the top events and the top event exist at time t if and only if
FTA is a well-known probabilistic approach in the
at least one of the n basic event occurs at time t, respectively.
domain of RAMS (reliability, availability, maintainability
Thus, the system unavailability is given by Eqs. (6) and (7),
and safety). FTA is a bottom-to-top (deductive) methodology
respectively.
for estimating the potential causes of accidents and failure
probabilities. FTA is centered about determining the causes i −n
of undesirable events, while referred as the top event. Fault A s (t ) = ∏ Ai = Pr( E1 ∩ E2 ... ∩ En ) = Pr( E1 )Pr( E2 )...Pr( En ) (6)
trees (FTs) are drawn at the top of the events. It the proceeds i −1

downward, dissecting the system in increasing details to

determine the root causes or combinations of causes of the i −n i −n
top event. Top events are usually considered as the major A s (t ) = 1 − ∏ (1 − Ai ) = Pr( E1 ∪ E2 ... ∪ En ) = 1 − ∏ (1 − Pr( Ei )
failure consequences, engendering serious safety hazards i −1 i −1

and the potential for significant economic loss. (7)

4 U. Safder et al. / Desalination and Water Treatment 109 (2018) 1–7
Fig. 2. Schematic representation of fault tree gates. (a) AND gate
and (b) OR gate.
3. Methodology of RO–FO desalination network
Probabilities of the basic events must be known in
advance in order to evaluate the failure probability of the top
event. The fuzzy set theory was used to get the probabilities
of the basic events in this study. The human error rate, the
expert’s subjective assessments with linguistic variables were
represented in fuzzy possibility scores (FPSs).
The evaluation data were expressed in linguistic terms
such as very low, low, fairly low, medium, fairly high,
high and very high. Fig. 3 represented the fuzzy number
of these linguistic values [19]. The fuzzy number is used
to handle imprecise information. There are many forms of
fuzzy numbers to represent the linguistic values. The tri-
angular and trapezoidal fuzzy numbers were used here to
denote failure possibilities. Transformation function was
used to convert the FPS into fuzzy failure rate (FFR) given
by Eqs. (8) and (9).
frequency of an error
FFR = (8)
total chance that an event having error
Swain and Guttmann [20] suggested that the error rate of
a routine human operation is 10–2 to 10–3 and the lower bound
of the error rate is 5 × 10–5.
1 / 10 , FPS ≠ 0 ,
k
FFR =  (9)
 0 , otherwise
Here, k = [(1 – FPS)/FPS]1/3 × a, where a = 2.301. The
constant a is appropriate estimation of failure rate of
desalination system. It can be realized that the range of
failure rate obtained by the fuzzy probability score corre-
sponds with the range of human error data suggestion by
Swain and Guttmann [20]. The value of k is taken from Lin
and Wang [19]. Lastly, unavailability of the integrated system
was estimated by combining the hardware failure rate and
the FPS of human error from the top event of FTA.
3.1. Integrated RO–FO system description
The schematic process flow diagram of the investigated
system is shown in Fig. 4 proposed by Kolluri et al. [8].
The integrated system produced freshwater in minimum
power consumption based on FO and RO principles.
The main system comprises two main stages: RO and
Fig. 3. The fuzzy number represents linguistic values [19].
Fig. 4. Schematic superstructure representation of RO–FO
integrated desalination system [8].
FO stage which are coupled mechanically. In the RO–FO
process, draw solution and seawater are pumped into
the FO membrane in an opposite direction of flow mode.
Freshwater will cross the FO membrane from the feed to
the draw solution side of the membrane in the direction
of the osmotic pressure gradient. After leaving the FO
membrane, now the diluted draw solution is input to RO
subsystem for freshwater production and osmotic agent
recycling and reuse. RO stage rejects salts and dissolved
contaminants that may have crossed the membrane from
the impaired water source. A second-stage FO process
can be implemented to dilute seawater before discharge
and to further concentrate and reduce the volume of the
impaired water stream. Most important, because brine of
RO subsystem is used as high salinity stream and seawater
is used as low salinity stream in FO subsystem, the energy
required for subsequent RO desalination of the diluted
saline water is reduced.
3.2. Hybrid FTA algorithm
A stepwise description of the hybrid FTA algorithm of
integrated RO–FO desalination system.
• Select the top event and construct FT logic diagram.
• Divide the elements of FT logic diagram into objective
probability analysis and subjective linguistic evaluation
of human performance and vague events.
• Select hardware failure-rate estimation from data bank.
• Conduct linguistic expressions for human performance
and vague events and transform into fuzzy number.
• Convert fuzzy numbers into FPSs and transform into
FFR.
• Integrate both failure-rate to estimate the failure-rate
of top event and unavailability or reliability of the
system.
 U. Safder et al. / Desalination and Water Treatment 109 (2018) 1–7
4. Results and discussion
The system availability was computed based on Oreda
report 2002 [21], recorded data and fuzzy set. A FT diagram
was drawn for the integrated system using AND and OR
gates as shown in Fig. 5. The output of these gates in terms of
unavailability was computed. The most important parameters
considered for estimating the unavailability of the RO–FO sys-
tem were semi-permeable membranes, filters, high-pressure
pumps (HPPs), pipeline leakages, instruments and controls.
Unavailability of the system due to power supply disruption
was not considered in the overall estimation of unavailabil-
ity. Since it is independent of the RO and FO technology.
The unavailability of the RO–FO system was estimated using
“AND” and “OR” gates as given in Eqs. (6) and (7).
4.1. Hybrid FTA failure rate calculation
Step 1: Divide the elements of fault-tree logic diagram
into objective probability analysis and fuzzy linguistic
Fig. 5. FT logic diagram of unavailability of integrated RO–FO system.
5
evaluation for human performance and vague events.
Events such as careless operation, noise, external leakages,
inadequate microbial control, monitoring and equipment
design were categorized as fuzzy subjective evaluation
events. Other events of system unavailability were cate-
gorized under objective probability analysis. The fuzzy
linguistic evaluation events were represented in shaded
blocks in Fig 5.
Step 2: Failure rate estimation was obtained by referring
the Oreda report 2002 [21]. The failure rate of corresponding
component faults shown in Table 1.
Step 3: For human performance and vague events, con-
ducted the linguistic assessment and assigned the scale
rating (such as very low, low, fairly low, medium, fairly
high, high and very high) to each fault based on expert
opinion.
Step 4: After obtaining the fuzzy number, convert
fuzzy numbers into FPSs. This was done by using Eqs. (10)
through (14).
6 U. Safder et al. / Desalination and Water Treatment 109 (2018) 1–7

Based on the calculated failure rate given in Tables 1 and 2,

 x , 0 ≤ x ≤ 1
fmax ( x) =  (10) the unavailability of the seawater pump was calculated from
0 , otherwise the top down event of FTA and using Eqs. (1) and (2).

1 − x , 0 ≤ x ≤ 1 Asw = 7.25 × 10 −4 (15)

fmin ( x) = 
 0 , otherwise (11)

FPS R ( M ) = sup[ f M ( x) ∧ fmax ( x)] (12) 4.2.2. Pretreatment system

x
The unavailability of the filter was zero:

FPS L ( M ) = sup[ f M ( x) ∧ fmin ( x)] (13)

x Afilter = 0 (16)

FPST ( M ) = [FPS R ( M ) + 1 − FPS L ( M )] / 2 (14) 4.2.3. FO stage

The fuzzy number and its corresponding failure
possibility scores of the fuzzy events shown in Table 2.
Step 5: Transform FPS into FFR, in order to integrate both
failure rate and estimate the unavailability of the integrated
RO–FO system using Eq. (9). The failure possibility scores of
the fuzzy events and its corresponding FFRs are shown in
Table 2.
Step 6: Calculate the unavailability of the subsystem of
the integrated network from estimated failure rates.
4.2. Calculation of reliability indices of integrated system
Based on the above failure rate data, the unavailability of
the subsystem was calculated in the following sections.
4.2.1. Seawater pump
Seawater pump supplied the entire feed to the FO system,
hence, when seawater pump failed, the system was down.
Since two trains of FO modules were taken in series,
the unreliability of the trains was to be multiplied to get the
unreliability of the FO subsystem. In addition, brine stream
of RO subsystem is pump to FO subsystem and considered
as high salinity stream in FO stage. From the FFR data on the
two trains, the unavailability was determined as follows:
AFO = 1.78 × 10 −3 (17)
4.2.4. RO stage
RO stage of the system was comprised of an
HPP, membrane module and stream to FO stage. The
unavailability of the each of the items of RO stage was
determined as follows:
ARO = 0.0132 (18)

Table 1 Table 3
Failure rate data for hybrid system Ranking of various subsystem of integrated RO–FO system

Failure events Failure rate/year Fault event Code Availability/h

1,111 2.45 × 10 –4 RO stage 13 0.9868

11,141 7.52 × 10–5 HPP 1,311 0.9870
11,142 6.51 × 10–5 FO stage 14 0.9982
11,151 7.06 × 10–5 Dosing pump 122 0.9986
11,152 2.26 × 10–4 Seawater intake 11 0.99927
11,211 7.19 × 10–4 RO membrane 1,312 0.99981
11,212 7.42 × 10–4 FO membrane 1,412 0.99983
1,311 0.01303 Filter 121 1.000

Table 2
Converting fuzzy numbers into fuzzy failure rate

Failure events/elements 1,112 1,113 11,143 113 1,312 1,412 1,122

FPSR (M) 0.275 0.384 0.237 0.196 0.315 0.305 0.176
1-FPSL (M) 0.055 0.064 0.0515 0.049 0.063 0.061 0.049
FPST (M) 0.165 0.224 0.1442 0.122 0.189 0.183 0.113
FFR 1.12 × 10–4 3.31 × 10–4 6.91 × 10–5 3.71 × 10–5 1.86 × 10–4 1.65 × 10–4 2.69 × 10–5
 U. Safder et al. / Desalination and Water Treatment 109 (2018) 1–7
4.2.5. Overall system reliability
The overall unavailability of the integrated RO–FO sys-
tem was determined by applying Eq. (15).
APlant = 1 − (1 − Asw ) × (1 − Apretreatment ) × (1 − AFO ) × (1 − ARO ) (19)
APlant = 0.0156 (20)
The availability of the integrated RO–FO desalination
system is approximately 0.9844/h. The availability of the sys-
tem appears to be high due to significant improvement in
integrated RO–FO desalination system. Table 3 presents the
most dominant fault paths based on their contribution to the
top event and sorting in increasing order of availability. These
fault events represent the potential areas for improvement. The
most likely cause is RO stage, which include RO membrane
module and brine stream to FO, followed by HPP. Thus, the
improvement of failure occurrence of the two events will have
the most significant contribution to enhance system safety and
reliability. While installing a new plant, engineers and plant
operators needed to keep the record of these two fault events in
priority level and safety interventions are proposed as follows:
• The plant operators should focus on the investigation of
the noise sources and external leakages of the HPP. They
can install an intelligent self-diagnose system to protect
the system from noises.
• The membrane module should maintained and perform
regular cleaning to protect from fouling.
This approach resolves some problems in traditional
method to estimate the failure rate of human performance.
The method is simple and intuitive in terms of evaluation
and computation. The analyst can give more reliable failure
assessment and can evaluate the failure events directly using
linguistic terms. The scarcity of objective and quantitative data
on human performance in integrated RO–FO desalination sys-
tem is a serious limitation. The failure probability represents
the best judgment based on expert’s opinion. However, this
study gives a more flexible and appropriate function struc-
ture to combine objective and subjective events to reflect the
diverse conditions in many industrial environments.
5. Conclusion
The RO is a major process for producing potable water
from seawater through desalination. The performance of any
desalination system depends on the failure behavior of its
subsystems. Since the hybrid RO–FO system was proposed as
power saving system, the reliability of the subsystems of RO
and FO is to be maintained at a high level by proper design
and selection of materials of these subsystems for operating
in the continuous plant. Thus, in this study, a hybrid FTA
method based on fuzzy set theory was used to directly esti-
mate the failure probability, the FFR and availability of the
system and provide the reliability information to the safety
engineers to easily evaluate the failure rate of the top event.
Since integrated system showed very high availability, this
integrated system can become viable to an alternative to single
RO desalination plants used globally. The design of integrated
View publication stats
7
RO–FO system is recommended because it has high perfor-
mance, low power consumption and economically feasible.
Acknowledgment
This work was supported by the National Research
Foundation of Korea (NRF) grant funded by the Korea
government (MSIP; No. 2017R1E1A1A03070713).
References
[1] E.H. Oelkers, J.G. Hering, C. Zhu, Water: is there a global crisis?
Elements, 7 (2011) 157–162.
[2] F. Calise, M.D. d’Accadia, A. Macaluso, L. Vanoli, A. Piacentino,
A novel solar-geothermal trigeneration system integrating
water desalination: design, dynamic simulation and economic
assessment, Energy, 115 (2016) 1533–1547.
[3] A. Bajpayee, T. Luo, A. Muto, G. Chen, Very low temperature
membrane-free desalination by directional solvent extraction,
Energy Environ. Sci., 4 (2011) 1672–1675.
[4] C. Li, S. Besarati, Y. Goswami, E. Stefanakos, H. Chen, Reverse
osmosis desalination driven by low temperature supercritical
organic Rankine cycle, Appl. Energy, 102 (2013) 1071–1080.
[5] T.Y. Cath, J.E. Drewes, C.D. Lundin, A Novel Hybrid Forward
Osmosis Process for Drinking Water Augmentation using
Impaired Water and Saline Water Sources, Water Research
Foundation, 2009.
[6] S. Sethi, S. Walker, J. Drewes, P. Xu, Existing & emerging
concentrate minimization and disposal practices for membrane
systems, Florida Water Resour. J., 58 (2006) 38–48.
[7] O. Bamaga, A. Yokochi, E. Beaudry, Application of forward
osmosis in pretreatment of seawater for small reverse osmosis
desalination units, Desal. Wat. Treat., 5 (2009) 183–191.
[8] S.S. Kolluri, I.J. Esfahani, P.S.N. Garikiparthy, C.K. Yoo,
A two-stage optimization approach for the synthesis of resource
conservation networks involving interception units, Chem.
Eng. Res. Des., 94 (2015) 52–71.
[9] M. Hajeeh, D. Chaudhuri, Reliability and availability assessment
of reverse osmosis, Desalination, 130 (2000) 185–192.
[10] I.I. Kutbi, Z.A. Sabri, A.A. Husseiny, Reliability analysis of
reverse osmosis plant, Desalination, 42 (1982) 291–313.
[11] K. Bourouni, Availability assessment of a reverse osmosis plant:
comparison between reliability block diagram and fault tree
analysis methods, Desalination, 313 (2013) 66–76.
[12] L.A. Zadeh, Fuzzy sets, Inf. Control, 8 (1965) 338–353.
[13] A. Der Kiureghian, O.D. Ditlevsen, J. Song, Availability,
reliability and downtime of systems with repairable
components, Reliab. Eng. Syst. Saf., 92 (2007) 231–242.
[14] A.K. Aggarwal, S. Kumar, V. Singh, Reliability and availability
analysis of the serial processes in skim milk powder system of a
dairy plant: a case study, Int. J. Ind. Syst. Eng., 22 (2016) 36–62.
[15] K.C. Kapur, M. Pecht, Reliability Engineering, John Wiley &
Sons, 2014.
[16] D. Bradt, Use reliability, availability and maintainability
techniques to optimize system operation, Hydrocarbon
Process., 76 (1997) 63–65.
[17] Y. Sun, Reliability Prediction of Complex Repairable Systems:
An Engineering Approach, PhD Thesis, School of Engineering
Systems, Faculty of Built Environment and Engineering,
Queensland University of Technology, 2006.
[18] M. Sánchez-Silva, G.-A. Klutke, Reliability and Life-Cycle
Analysis of Deteriorating Systems, Springer International
Publishing, 2016.
[19] C.-T. Lin, M.-J.J. Wang, Hybrid fault tree analysis using fuzzy
sets, Reliab. Eng. Syst. Saf., 58 (1997) 205–213.
[20] A.D. Swain, H.E. Guttmann, Handbook of Human-Reliability
Analysis with Emphasis on Nuclear Power Plant Applications.
Final Report, Sandia National Labs., Albuquerque, NM, USA, 1983.
[21] H.A. Sandtorv, P. Hokstad, D.W. Thompson, Practical
experience with a data collection project: the OREDA project,
Reliab. Eng. Syst. Saf., 51 (1996) 159–167.